Open AI introduced GitHub interlocking functions in ‘Deep Research’, a man-made intelligence agent, to greatly strengthen code evaluation capability and utilization.
Open AI announced on August 8 (local time) that it can be provided for a number of days to ‘Chat GPT Plus’, ‘Pro’, and ‘Team’ users by adding the GitHub Code Base Evaluation to Deep Research on August 8 (local time). Chat GPT Enterprise and Edu will soon support.
This feature is the primary integrated tool unveiled under the name ‘Connector’. Consequently, developers can analyze public code storage through the chat GPT, understand technical documents, and receive answers to specific code questions.
“Many users wish to connect the deep research of the chat GPT with not only the online but additionally the inner resources in addition to the online,” said Nate Gonzalez Open.
▲ Q & A of codebase ▲ Reconstruction of product specifications as technology tasks ▲ Code structure and pattern summary ▲ grasp the API implementation method using actual code examples.
Nonetheless, he still emphasized that hallucinations exist as a consequence of the character of AI, and that skilled engineers are still vital.
By way of security, the organization’s settings are observed and designed to be viewed only within the range of GitHub content allowed by the user and the explicitly shared codebase.
The introduction of this connector reflects the strategy of open AI to attach the chat GPT with an external platform. Earlier, Antropic introduced ‘Integrations’, an external app connection function to ‘Claude’, and Open AI also caught up with the present plug-in system and switched to a custom GPT.
At the identical time, it is usually accelerating the advancement of the event tool. Recently, it has launched an open source coding tool for terminals, Codex CLI, and the Desktop Chat GPT app has also strengthened the code recognition function along side major developer tools. At the identical time, the acquisition of AI coding startup ‘Windsurf’ was acquired for $ 3 billion.
The open AI also expanded the high-quality adjustment function for AI model customization. ‘O4-Mini’ has been able to bolster the reinforcement high-quality adjustment using the scoring method by work, and ‘GPT-4.1 Nano’ also supports high-quality adjustments to all paid developers.
Nonetheless, O4-mini can only be utilized by authentication organizations. Open AI has been restricting access to advanced models and a number of the developer functions since April, which is a measure to forestall abuse.
By Park Chan, reporter cpark@aitimes.com
